Understanding Society has a very complex survey structure and this needs to be taken into account in any analysis in order to infer results to a population. An easy way to take the structure into account correctly is by using weights alongside strata and clustering adjustments. This course aims to explain what weights take into account in Understanding Society. First, we will describe how weights are created. We will then provide guidance on how to select the best weight for a range of analyses. The course will then expand on how to create a tailored weight for a specific analysis, when this can make a difference and will be worth the effort, and will walk you step by step on creating such weights in a hands-on session.
This course is aimed at new users of Understanding Society, as well as those who have so far only made use of simpler aspects of the data. It aims to guide the user through the complexities of using this data for cross-sectional and longitudinal analysis, and ensure that they can make effective use of the data for their own research projects.
Introduce the basic principles of causal modelling (potential outcomes, graphs, causal effects) while emphasising the key role of design and assumptions in obtaining robust estimates.
Introduce the basic principles of machine learning and the use of machine learning methods to do causal inference (e.g. methods stemming from domain adaptation and propensity scores).
Show how to implement these techniques for causal analysis and interpret the results in illustrative examples.
This introductory course will show what panel data CAN and CANNOT do to facilitate the study of causal effects. There is a strong emphasis on applied work using linear models and panel data econometrics methods will be illustrated using data from Understanding Society. The software used for the practical examples is Stata. Familiarity with the basic syntax and structure of Stata is required.
This course is aimed at new users of Understanding Society, as well as those who have so far only made use of simpler aspects of the data. It aims to guide the user through the complexities of using this data for cross-sectional and longitudinal analysis, and ensure that they can make effective use of the data for their own research projects.
his is a two day course which will be taught over four mornings.
The fact that correlation does not equate to causation is so well known that it has become a popular saying in itself. Yet the way that quantitative analysis is discussed in much popular and political discourse, as well as interpreted by many social scientists, fails to take issues surrounding causality fully into account. This may be because randomized control experiments, widely understood as the most defensible method of establishing causality, are frequently impossible or unethical to conduct in social science settings.
Analysts thus have to work with observational data, which often miss information crucial for making causal interpretations of statistical associations. However, under some circumstances and subject to specific assumptions, one can interpret estimated associations as casual with substantially higher confidence. This course deals with methods that can be used under such circumstances and subject to the specific assumptions. The course offers practical skills in implementing these methods and the theoretical skills needed to understand and discuss evidence from them.
This course is aimed at new users of Understanding Society, as well as those who have so far only made use of simpler aspects of the data. It aims to guide the user through the complexities of using this data for cross-sectional and longitudinal analysis, and ensure that they can make effective use of the data for their own research projects.
We provide the first causal estimates of impacts of individual job displacement of men and women on the risk of domestic violence and also the first estimates of whether unemployment benefits mitigate these impacts. Estimating this confluence of three parameters on a given sample places us in a strong position to illuminate the underlying mechanisms. Using data on the 2 million domestic violence cases brought to criminal courts in Brazil during 2009-2018, we identify the defendant and plaintiff in longitudinal employment registers. Leveraging mass layoffs for identification, we find that both male and female job loss, independently, lead to a large, pervasive increase in domestic violence. Exploiting a discontinuity in unemployment insurance eligibility, we find that it does not ameliorate while benefits are being paid, and that eligible men are more likely to commit domestic violence than ineligible men once benefits expire. Our findings are consistent with job loss constituting a negative shock to income and identity and a positive shock to time. Unemployment benefits tend to offset the income shock but reinforce the time shock.
Dr Amy Clair’s work with Adam Tinson for the Health Foundation thinktank shows housing problems are likely a significant component of the ‘syndemic’ that has led to greater risks of Covid-19 infection and serious complications for certain social groups
Dr Ayse Guveli from our ESRC Research Centre on Micro-Social Change describes her findings on the impact of periods of home schooling on inequalities in this article for The Conversation
The Covid-19 pandemic has exposed and amplified the strengths and weakness in our society and our economy. Register for our programme of topical online discussions during the week of 25 - 29 January 2021
Increase in mental health disorders during the first lockdown appears to have been linked to people feeling lonely, having Covid symptoms, working from home, caring for children and struggling financially
The goal of this workshop is to bring together researchers using subjective expectations in their work and to foster and stimulate the exchange of ideas. The workshop is the sixth in a series of workshops on Subjective Expectations. Previous workshops have been held at Université Laval in Québec city (2009 and 2011), the University of Essex (2014), New York Federal Reserve Bank (2016), and CESifo in Munich (2018). (Two half days, programme under preparation).
Programme and keynote speakers:
There will be three themed sessions based on available research data and each followed
by group discussions and a plenary:
1. Impact - Impact of Covid-19 on society, health and higher education - Dr Adrian James,
President Royal College of Psychiatrists
2. Inequalities (Title TBC) - Sarah Hughes, CEO, Centre for Mental Health
3. Innovation - The surge in domestic violence during the pandemic: Can creative/
innovative research illuminate its causes? - Professor Sonia Bhalotra, Professor of
Economics, University of Essex
This webinar will launch the final report of Birgitta Rabe and Angus Holford’s Nuffield Foundation-funded project on the impacts of the Universal Infant Free School Meals (UIFSM) policy, and bring together researchers, practitioners and policymakers to debate how school or public provision of food to children should be developed in the UK
Full report by Dr Birgitta Rabe and Dr Angus Holford on our Nuffield Foundation-funded study into the impact of the controversial government policy published today, with important findings on obesity
The milestone of Wave 10 data released from our world-leading UK household longitudinal study, with extra modules including sleep quality, pensioner deprivation, commuting, voluntary work and voting in the last General Election
This course is aimed at new users of Understanding Society, as well as those who have so far only made use of simpler aspects of the data. It aims to guide the user through the complexities of using this data for cross-sectional and longitudinal analysis, and ensure that they can make effective use of the data for their own research projects.
The course is free to attend and includes lunch and refreshments.
The day-long course online will introduce you to various empirical, quantitative methods that can be used to estimate the impact of a specific policy intervention. These methods can be referred to as “programme evaluation”, “impact assessment”, “causal estimation” or “impact evaluation”. The course assumes basic statistical concepts (mean, median, correlation, expected value, statistical significance and confidence intervals), and algebra is optional. It does not teach participants how to implement any of these methods using statistical software.
Even before the pandemic, pay insecurity was rising in Britain driven by the boom in the ‘gig’ economy, rising self-employment and growth of zero-hours and other types of jobs that guarantee no or only limited earnings.